Author Topic: RPS with from 0-30V and Current variable up to 3 A design Circuit Please....  (Read 8166 times)

0 Members and 1 Guest are viewing this topic.

Offline techguruTopic starter

  • Regular Contributor
  • *
  • Posts: 83
  • Country: in
Dear all,
 

                   For past three month i have been designing a variable power supply which gives 0-30V and Current variable up to 3 ampere from zero. can some can give working Schematic so that i would understand and complete my project. thank you all..
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
https://www.eevblog.com/forum/beginners/diy-0-30v-0-3a-again/  See Reply #7

Most of these projects fail to meet the specifications.  Some can't get down to 0V because they don't have a negative supply for the op amps.  Some can't handle low voltage with high current because there is too much voltage drop across the output transistors.  None have transformers with taps to solve this issue. 

The Internet is full of circuits that don't work.  If there was one that did, it should be posted as a 'sticky'.
 
The following users thanked this post: techguru

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
30 volts at 3 amps is a tough row to hoe without a lot of extra complexity.  90 watts worst case requires multiple transistors and 30 volts peak is marginal for common operational amplifiers which have an absolute maximum voltage rating of 36 volts.  20 volts at 1 amp is much easier.
 
The following users thanked this post: techguru

Offline techguruTopic starter

  • Regular Contributor
  • *
  • Posts: 83
  • Country: in
Hi to all,
            The above said opamp Vcc problem is true.
What are the opamp whose Vcc greater than 40V and above in single supply operation. So that I could use it in my circuit.thank you all.
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7549
  • Country: 00
  • +++ ATH1
                   For past three month i have been designing a variable power supply which gives 0-30V ....

3 months ? Its been almost a year Gurumurthy ... (this guy previous nick).

The same guy that keeps haunting this forum for his 723 based PSU, as he keep posting another new thread "AGAIN AND AGAIN" if he does not have the component in his drawer as the circuit needs, and keeps nagging until someone give him a circuit that can do 30V 3A 723 based PSU with components that he owns.  :palm:

Reference -> https://www.eevblog.com/forum/projects/how-to-calculate-potentiometer-value/msg1237728/#msg1237728

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
30 volts at 3 amps is a tough row to hoe without a lot of extra complexity.  90 watts worst case requires multiple transistors and 30 volts peak is marginal for common operational amplifiers which have an absolute maximum voltage rating of 36 volts.  20 volts at 1 amp is much easier.

since neither a zero ripple voltage unregulated DC source nor a zero drop voltage power device exist, in the real world, the worst case would require - at the very least - 100W (may be even 110-120W) of power dissipation  :D
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
What are the opamp whose Vcc greater than 40V and above in single supply operation. So that I could use it in my circuit.thank you all.

If you do not mind paying a premium, the LT1006/LT1013/LT1014 (single/dual/quad) operational amplifiers are single supply and 44 volts replacements for the common 358/324.  There are also newer low cost parts like the MC33171 and MC34171 series which I have never used.

There are some higher voltage specialty parts but an alternative is to design the control circuits so they either float using the output voltage as ground or have a separate floating supply from another transformer winding.  If the low voltage floating supply is bipolar, then any operational amplifier can be used.
 
The following users thanked this post: techguru

Offline techguruTopic starter

  • Regular Contributor
  • *
  • Posts: 83
  • Country: in
Is it possible to construct my requirements with opamp and discrete components?
give me some idea to construct the circuit.


Kindly guide me in this direction.. thank you all...
« Last Edit: January 07, 2019, 04:43:27 pm by techguru »
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Yes, almost all audio power amplifiers are like a big op-amp. You will need only the transconductance part driving the output Darlington.
There will be complications like adding current regulation and making it regulate down to zero, likely requiring a - rail. You will end up with a monster like most other designs out there.
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 
The following users thanked this post: techguru

Offline techguruTopic starter

  • Regular Contributor
  • *
  • Posts: 83
  • Country: in
If any other possibility are there?

Like Mr.bravo said I am trying this project for more than a year.now I came to know that linear mode that is class A mode of design is intractable.

You all guide in design class s amplifier (smps), where my efficiency is nearly 100 percent. Buck topology.I think I am moving in right direction.help me in this regard.
« Last Edit: January 08, 2019, 12:30:12 pm by techguru »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
I've never heard of a class S amplifier don't you mean class D?

Why do you need a 0 to 30V, 0 to 3A  PSU for an audio amplifier? I presume you want a general purpose bench power supply, rather than a dedicated audio amplifier PSU. A decent audio amplifier design shouldn't need a tightly regulated power supply. It should be designed to reject changes in the power supply voltage, without passing them through to the speakers.

The problem with a constant voltage and constant current PSU is Ohm's law states that it can't be in both modes simultaneously. When the load resistance is above a certain point which would exceed the output voltage setting, it will be in constant voltage mode, then when the load resistance drops below the point when the current limit would be exceeded, it will switch to constant voltage mode. The problem is switching from constant current to constant voltage modes and back with minimal oscillation and current/voltage over/undershoot. Oscillation can be avoided, but it's actually impossible to prevent over/undershoot.

Here's a simple constant VI regulator I lashed up in LTSpice. U1 needs to work with its inputs down to 0V and U2 with its inputs up to +V. V2 and V3 set the voltage and current limits respectively. D1 turns on when the regulator current limits, otherwise it's off. RL1/2/3 test the transient voltage and current response. RL1 applies a step load of 500mA, for 10ms. RL2 takes it into constant current mode and back again. RL3 as good as short circuits the power supply, when it's in the constant current mode.

C2, C2, R8, R9 and R10 form phase compensation networks which reduce the oscillation.

The transient response isn't great. Adding a capacitor across the output will cut down on the voltage over/undershot, at the expense of current overshoot and adding an inductor will improve the current mode, at the expense of the voltage mode.

« Last Edit: January 08, 2019, 01:36:31 pm by Zero999 »
 

Offline techguruTopic starter

  • Regular Contributor
  • *
  • Posts: 83
  • Country: in
Sorry it is class d amplifier. Thanks a lot. I will try to understand the schematic.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
If any other possibility are there?

Like Mr.bravo said I am trying this project for more than a year.now I came to know that linear mode that is class A mode of design is intractable.

You all guide in design class s amplifier (smps), where my efficiency is nearly 100 percent. Buck topology.I think I am moving in right direction.help me in this regard.

If there was a circuit out in the wild that would actually deliver 0-30V and 0-3A, it would be all over this forum.  It would be a sticky on every forum.  Everybody on the Internet would know where to find it.  Instead of dozens of replies, the matter would be settled with a simple link and the comment to "build this".

Alas, it's just a lot harder than it looks.  Any design without 3 or 4 parallel pass transistors is probably not going to deliver 3A at 0.1V without transformer taps.  High voltage op amp selection is a challenge.  There's a reason that 'real' bench power supplies cost so much money.  There's a lot of parts!

If the power supply is for a fixed load, like an audio amplifier, the 0-30V 0-3A spec doesn't make sense.  If it's for a bench PS then why in the world do you need anywhere near 3A?  For that matter, why 30V?  A 1A adjustable supply is perfectly reasonable for the bench and, in fact, adjustable voltage isn't a real requirement either.  5V, +-15V and, maybe, +12V fixed 1A supplies will probably cover just about every conceivable situation that is 'general purpose'.  Sure, vacuum tube circuits clearly need other voltages but that arena is covered by dedicated fixed supplies.

My view on the 0-30V 0-3A:  Somebody wrote it down once and now it is the assumed standard in the industry.  In fact, there are few DIY projects that succeed.  Try audioguru's project and if it works out, it will be one of the few designs that actually delivers.  But don't substitute parts!  See Reply #6 here:

https://www.eevblog.com/forum/beginners/diy-0-30v-0-3a-again/

As a reference point, the Rigol DP832 PS costs about $500.  It has 3 outputs that will presumably deliver up to 3A each, 2 at 0-30V and 1 at 0-5V.  But it costs $500 and there's a lot of reasons why.

There are a lot of cheap power supplies - they are all over the Internet.

https://www.amazon.com/Lab-Power-Supplies/b?ie=UTF8&node=318022011


« Last Edit: January 08, 2019, 06:06:35 pm by rstofer »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
The first example below is a design which almost meets your requirements.  I would not implement it as shown because it uses LM395 integrated power transistors but it could be modified to use normal transistors.  It could also be modified to work without a negative supply.  Notable features are very low output capacitance and fast transitioning between constant current and voltage modes.

The second example below from the Tektronix PS501 is closer to what I would do.  Notably:

1. The current sensing is on the high side of the output making it more accurate.  The current shunt in series with the output can be used to make the regulation loop more stable although that was not done in the PS501.
2. For a high power supply, I would replace the output transistor with an integrated regulator like a 317 driving a set of big power transistors.  This allows the integrated regulator to protect the output transistors however it also means that a negative supply is required to allow the output to get down to zero volts.
3. I would add clamping to the error amplifiers to reduce transition time between voltage and current mode.
 
The following users thanked this post: techguru

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
The first example below is a design which almost meets your requirements.  I would not implement it as shown because it uses LM395 integrated power transistors but it could be modified to use normal transistors.  It could also be modified to work without a negative supply.  Notable features are very low output capacitance and fast transitioning between constant current and voltage modes.


That National LB-28 design seems pretty nice.  Minimal components, reasonable specs and the LM395 is still available albeit expensive ($4).  Still, a quality supply just can't be built for a buck ninety five.

Changing to normal transistors will be a challenge for a newcomer.  The LM395s only take 10 uA of base current (3 uA typical) so the op amp can easily drive several.  In addition, the LM395 has a lot of built-in protection, well beyond that of a 2N3055.

I would be tempted to build the LB-28 as shown.

http://www.ti.com/lit/ds/symlink/lm395.pdf

It's still going to take a pretty beefy heatsink!

ETA:  The LM395 also has a 0.1 Ohm ballast resistor to aid in current sharing across multiple devices.

« Last Edit: January 08, 2019, 10:41:24 pm by rstofer »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
That National LB-28 design seems pretty nice.

Although as shown it requires a -5 to -10 volt negative supply.

Quote
Minimal components, reasonable specs and the LM395 is still available albeit expensive ($4).  Still, a quality supply just can't be built for a buck ninety five.

Check the schematic carefully; the LB-28 design uses 7 LM395s in parallel for 8 total although new big TO-3 power transistors are not much less expensive.

Quote
Changing to normal transistors will be a challenge for a newcomer.  The LM395s only take 10 uA of base current (3 uA typical) so the op amp can easily drive several.  In addition, the LM395 has a lot of built-in protection, well beyond that of a 2N3055.

I brought it up more as an example to study.  What I might do is use one LM395 to drive a bunch of big current boosting transistors but like you said, this would be a challenge for a newcomer.
 
The following users thanked this post: techguru

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
If any other possibility are there?

Like Mr.bravo said I am trying this project for more than a year.now I came to know that linear mode that is class A mode of design is intractable.

You all guide in design class s amplifier (smps), where my efficiency is nearly 100 percent. Buck topology.I think I am moving in right direction.help me in this regard.

If there was a circuit out in the wild that would actually deliver 0-30V and 0-3A, it would be all over this forum.  It would be a sticky on every forum.  Everybody on the Internet would know where to find it.  Instead of dozens of replies, the matter would be settled with a simple link and the comment to "build this".

I do not think there is a lack of such schematic diagrams neither that it would be more difficult to design than e.g. a 0-15V 1A PSU.
4 large heatsinks and 8 TO3 BJTs can work for more than 120W of wasted power even without the help of fans.
It just does make little sense and would be more expensive than using a suitable multi-tap transformer based PSU.
 
The following users thanked this post: techguru

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
If any other possibility are there?

Like Mr.bravo said I am trying this project for more than a year.now I came to know that linear mode that is class A mode of design is intractable.

You all guide in design class s amplifier (smps), where my efficiency is nearly 100 percent. Buck topology.I think I am moving in right direction.help me in this regard.

If there was a circuit out in the wild that would actually deliver 0-30V and 0-3A, it would be all over this forum.  It would be a sticky on every forum.  Everybody on the Internet would know where to find it.  Instead of dozens of replies, the matter would be settled with a simple link and the comment to "build this".

I do not think there is a lack of such schematic diagrams neither that it would be more difficult to design than e.g. a 0-15V 1A PSU.
4 large heatsinks and 8 TO3 BJTs can work for more than 120W of wasted power even without the help of fans.
It just does make little sense and would be more expensive than using a suitable multi-tap transformer based PSU.
Go on, post your design. . .

He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.
 
The following users thanked this post: techguru

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.

So use a separate 40 volt low power regulator for the control circuits; it might be as simple as a zener shunt regulator.  Or configure the control circuits in a floating configuration with a low power negative regulator following the output voltage; this is more suitable for even higher voltage supplies.  With a floating design, low voltage operational amplifiers can be used.  The alternative to a floating design is to add high voltage level shifting between the operational amplifier and pass transistors which is a whole different set of problems; most new designers go this route.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.

So use a separate 40 volt low power regulator for the control circuits; it might be as simple as a zener shunt regulator.  Or configure the control circuits in a floating configuration with a low power negative regulator following the output voltage; this is more suitable for even higher voltage supplies.  With a floating design, low voltage operational amplifiers can be used.  The alternative to a floating design is to add high voltage level shifting between the operational amplifier and pass transistors which is a whole different set of problems; most new designers go this route.
My point was that none of that is necessary with a 0 to 15V supply.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.

So use a separate 40 volt low power regulator for the control circuits; it might be as simple as a zener shunt regulator.  Or configure the control circuits in a floating configuration with a low power negative regulator following the output voltage; this is more suitable for even higher voltage supplies.  With a floating design, low voltage operational amplifiers can be used.  The alternative to a floating design is to add high voltage level shifting between the operational amplifier and pass transistors which is a whole different set of problems; most new designers go this route.
My point was that none of that is necessary with a 0 to 15V supply.

I think this is the point I was trying to make.  The 0-30V 0-3A spec is nice to read but relatively difficult to achieve for a number of reasons.  Further, it isn't really necessary for the vast majority of bench type projects.

It isn't just finding a schematic, it's also having to make a PCB and, more important, lay it out properly.  Fewer parts is easier.

I wonder if the cheap supplies can actually deliver 3A at 0.1V.  In fact, I wonder how my DP832A would react to that.  The good news is that I don't have an electronic load to test it.  I'm not sure how it would come out.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
I do not think there is a lack of such schematic diagrams neither that it would be more difficult to design than e.g. a 0-15V 1A PSU.
4 large heatsinks and 8 TO3 BJTs can work for more than 120W of wasted power even without the help of fans.
It just does make little sense and would be more expensive than using a suitable multi-tap transformer based PSU.
Go on, post your design. . .

He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.

You must be kidding... nothing easier than that.  ;D

I made this from a quick copy & paste from other .asc files so there are likely some problems although it works flawlessly in the simulator.

BTW to play with it you need TL431 and BD140 models.

« Last Edit: January 09, 2019, 07:06:03 pm by not1xor1 »
 
The following users thanked this post: techguru

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
I do not think there is a lack of such schematic diagrams neither that it would be more difficult to design than e.g. a 0-15V 1A PSU.
4 large heatsinks and 8 TO3 BJTs can work for more than 120W of wasted power even without the help of fans.
It just does make little sense and would be more expensive than using a suitable multi-tap transformer based PSU.
Go on, post your design. . .

He's right. It's not so much the current, but the voltage which is the problem. At 15V, an ordinary op-amp can be run from the unfiltered rectifier voltage. With 30V something more elaborate is required. The absolute maximum voltage rating of most standard op-amps is under 44V, which would be an AC voltage of just over 30V on the transformer secondary. A lower voltage transformer could be used, but that doesn't leave much headroom for voltage ripple.

You must be kidding... nothing easier than that.  ;D

I made this from a quick copy & paste from other .asc files so there are likely some problems although it works flawlessly in the simulator.

BTW to play with it you need TL431 and BD140 models.
Many of those models are not in the standard LTSpice install, but I got it working eventually.

The only issue which immediately caught my attention is the rectifier diodes are underspecified, but that's easy to fix. No doubt it will work, after a bit of tweaking, but it's not a beginner project, like a PSU with a lower output voltage would be.

Please pay a bit more attention to the schematic layout: it's very hard to follow!
 
The following users thanked this post: techguru

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
[...]
I made this from a quick copy & paste from other .asc files so there are likely some problems although it works flawlessly in the simulator.

BTW to play with it you need TL431 and BD140 models.
Many of those models are not in the standard LTSpice install, but I got it working eventually.

The only issue which immediately caught my attention is the rectifier diodes are underspecified, but that's easy to fix. No doubt it will work, after a bit of tweaking, but it's not a beginner project, like a PSU with a lower output voltage would be.

Please pay a bit more attention to the schematic layout: it's very hard to follow!

Yes I forgot a few other parts. I was in a hurry... otherwise I would have embedded at least the BJT models.
The schematic is a bit messy. I should have split at least the AC-DC + Vcc/Vee parts from the control circuit.

I'm in the process of designing a strip-board layout for a first low power version of the circuit (0-18V 1A) although I'm not sure when I'll be able to build it.
 

Offline JohnPen

  • Regular Contributor
  • *
  • Posts: 240
  • Country: gb
I built  a PSU quite a few years ago from a circuit, I believe in the Wireless World you can tell how many years ago, that provided 2 x 22 volts at 2A.  It used separated windings on a 100 VA TFR to power the isolated sub PSU outputs.  Each sub PSU used a 78M12, directly from the rectified output, to power the control circuits consisting of 2 CA3140 (fet opamps).  The final regulator transistor was the good old 2N3055.   The voltage output was adjustable over ~ 0.5v to 22v.  It also has a adjustable current limit range ~20ma to just over 2A.  A separate overwind of a few turns on the TFR provided an extra low voltage AC feed to provide for a 7805 regulator to deliver a fixed 5v supply.  I felt at the time that it provided all the voltage/current/protection I would need for hobby use which included building a home computer.  Overall I found this PSU provided everything I wanted and only built one off dedicated supplies for projects such as Audio amplifiers that need higher voltages/current etc.  Out of curiousity I checked today what the lowest output voltage on one of the outputs could provide at ~2A and found it to be 0.5 volts with rapidly warming heatsink. :)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf